[GRASSLIST:10503] Re: [GRASS5] gis manager 2 rc3 - bug fixes and updates

[Michael Barton]

I just posted a set of updates to gism2 to the cvs and my website.

These are mostly bug/annoyance fixes. Maciek has been helpful in identifying
these, so I will list them below. I want to freeze features so that gism can
be tested and remaining bugs eliminated.

IMPORTANT NOTE. There has been a recent change in the way the monitorcap
file is created that is causing problems in starting any command that
requires an xmon display from the GIS Manager. I think it is an error in the
makefile for XDRIVER, but I haven't messed with this.

Monitorcap is being created with a double set of xmon definitions.
Currently, you will need to edit this file to delete the second set of
definitions for xmon (x0-x6). Hopefully, this can be corrected.

Michael
______________________________
Michael Barton, Professor of Anthropology
School of Human Evolution and Social Change
Arizona State University
Tempe, AZ  85287-2402
USA

voice: 480-965-6262; fax: 480-965-7671
www: http://www.public.asu.edu/~cmbarton


> From: Maciek Sieczka <werchowyna at epf.pl>
> Date: Tue, 21 Feb 2006 08:58:25 +0100
> To: <michael.barton at asu.edu>
> Cc: <grass5 at grass.itc.it>, <glynn at gclements.plus.com>
> Subject: Re: [GRASSLIST:10435] Re: [GRASS5] gis manager 2 rc3 - bug fixes and
> updates
> 
> Michael,
> 
> Before gis.m stopped working for me due to recent tcl/tk 8.3
> incompatibility, I have noted down following remaining issues.
> 
> CCing Glynn as some issues are display related.
> 
> 1. Increase the width of main gis.m, a bit. Currently the properties of
> a vector layer ("Display vector maps") doens't fit (few pixels) thus
> the horizontal slider bar is there - actually no real need for it.
> 

Done

> 
> 
> 2. g.region zombies are gone. d.mon ones still bite. You asked me once
> how to find them. Home-made zombie detector: ps -A -f | grep defunct.
> 

I don't know how to zap these for d.mon. Any suggestions?

> 
> 
> 3. When reducing the gis.m window size, the buttons in the bottom of the
> window ("run" "clear" etc.) get hidden. When I enlarge the window back
> they come back but a huge amount of space becomes wasted in the window.
> Please do some resizing the window for and back to see what I mean.
> Could the buttons be always left visible? And the window space not
> wasted?
> 

Fixed in last Friday's release. These no longer get hidden.

> 
> 
> 4. The "Digitize map" button doesn't work. In case of rasters something
> pops up but dissapears in a flash. For vectors an error is issued:
> 
> child process exited abnormally
> child process exited abnormally
>
snip snip

> And terminal reads:
> 
> using default visual which is TrueColor
> ncolors: 16777216
> Graphics driver [x0] started
> Socket is already in use or not accepting connections.
> Use d.mon to select a monitor
> Problem selecting x0. Will try once more
> Socket is already in use or not accepting connections.
> Use d.mon to select a monitor
> 
> 

This is fixed IF monitorcap is edited as I indicated above. The same problem
affects other commands that need to open an xmon. The GIS Manager will do
this automatically if monitorcap is changed.

> 
> 5. The current tool should be automatically de-selected when chossing
> another tool. Currently it's not:
> 
> a) Use zoom, unzoom or pan tool.
> b) Don't quit it.
> c) Use query tool.
> d) See how the tool you used in 1. still affects the display (moves,
> resizes etc.)
> 

Implemented in this release. This is the only new feature introduced here.
Map display tools now behave like radiobuttons.

> 
> 
> 6. gis.m will fail to display a raster of name identical to another in
> different mapsets; and the "Map display" window will become broken after
> that:
> 
> WARNING: 'cell/x' was found in more mapsets (also found in topo).
> WARNING: 'cell/x' was found in more mapsets (also found in topo).

Due to the information put out by d.rast I think. If the map is listed
properly for non-current mapsets (<file>@<mapset>) this probably won't
happen. But I haven't had a chance to look into this yet.

> 
> 
> 7. The "Show data" button will not work for vectors outside the current
> mapset, at least in case of DBF driver. This is due to you have specify
> the path to a database for db.select implicitely at present:
> 
> a) I'm in mapset "topo".
> b) To display data of a DBF table of a vector "cieki" in mapset "melio"
> I have to use:
> 
> db.select cieki database='$GISDBASE/$LOCATION_NAME/melio/dbf/'
> 
> The way gis.m and d.m currently is doing it is:
> 
> db.select table=cieki at melio
> 
> Which leads to an error:
> 
> dbmi: Protocol error
> dbmi: Protocol error
> 
> There is a thread open for this issue in the bugtracker. Markus says it
> is not a bug. I'll close it as soon as I'm convinced. Currently I don't
> quite understand why "db.select table=cieki at melio" should not work.
> https://intevation.de/rt/webrt?serial_num=4083
> 
> On top of it, I'm not sure if using db.select for the purpose you are
> using it is a good idea - the name of a datatable happens to be
> different than the vector file name. The way "Show data" is now will
> work for sure only for vectors created by Grass commands which
> automatically create a table of the same name as the vector file. But
> would not in case of multi layer vector files (e.g. v.clean output),
> when the table names for each layer differ from the vector file name by
> a _<number> suffix.
> 
> Does anyody know what else command could be used then instead of
> db.select to avoid those problems?

This now fixed for the GIS Manager. db.select will use the table connected
to the vector and specified in the layer setting, regardless of where it is
located or what it is named. It will get the relevant settings from
v.db.connect -g

> 
> 
> 
> 8. The menu Xtns contains an active button "Add extensions here",
> clicking which leads to an error for obvious reason - obvious for
> somewhat experienced Grass user, newcomers might be puzzled. Could this
> button be replaced with a plain text or removed?

This is for future expansion. I'm working with Benjamin Ducke on this and it
should be available soon. But we need to get the bugs out of this release
before moving ahead and implementing this.

> 
> 
> 
> 9. Wish: make mouse wheel working in all the d.m/gis.m windows. This
> issue was mentioned on the list before, few moths ago, and the
> conclussion was it would be not doable for then. Any progress/chances
> in this regard currently?

I don't know.

> 
> 
> 
> 10. Zooming in is not working properly:
> 
> a) v.in.region output=frame
> b) add the vector 'frame' to gis.m layer tree
> c) g.region vect=frame
> d) "Zoom to current region" in "Map display"
> e) display
> f) you should see the whole vector 'frame', but you don't
> 
> This reminds me of https://intevation.de/rt/webrt?serial_num=3613.
> Still open.

Still working on this.

> 
> Maciek
> 
> --------------------
> W polskim Internecie s? setki milion?w stron. My przekazujemy Tobie tylko
> najlepsze z nich!
> http://katalog.panoramainternetu.pl/
>